ADHD is a psychiatric disorder affecting between 3% and 5% of the general population (American Psychiatric Association [APA], 2000; Kessler et al., 2006). The defining features of the disorder are inattention, impulsivity, and hyperactivity that impair functioning. There are three subtypes of ADHD: the predominately hyperactive-impulsive type, the predominantly inattentive type, and the combined type (APA, 2000). ADHD is most diagnosed during elementary school yet continues to persist into adolescence and adulthood for the majority (50%-70%) of individuals (Barkley, Murphy, & Fischer, 2007; Biederman et al., 1993, 2007; Faraone, Biederman, & Mick, 2006; Kessler et al., 2006). Like in children, ADHD in adolescents and adults is most often characterized by defining symptoms of hyperactivity/impulsivity and inattention (Faraone et al., 2006).

ADHD is very often comorbid with other forms of psychopathology. Comorbid anxiety, mood, disruptive behavior disorders, and substance use disorders are commonly reported in pediatric, adolescent, and adult ADHD (Barkley et al., 2007; Barkley, Murphy, & Kwasnik, 1996; Biederman et al., 1993; Borland & Heckman, 1976; Group, 1999; Heiligenstein, Conyers, Berns, & Miller, 1998; Kessler et al., 2006; Murphy & Barkley, 1996; Shekim, Asarnow, Hess, Zaucha, & Wheeler, 1990; Thomson et al., 2005). In addition, ADHD diagnoses are often accompanied by increased risk for learning disabilities as well as impaired social relationships (Busch et al., 2002; Fischer, Barkley, Smallish, & Fletcher, 2002; Mannuzza, Klein, Bessler, Malloy, & Hynes, 1997).

In the majority of children with ADHD (65%), impairing ADHD symptoms continue into adolescence (Ingram, Hechtman, & Morgenstern, 1999). The burden of illness arising from the persistence of ADHD is varied, pervasive, and significant in personal, familial, and economic costs (Fletcher & Wolfe, 2009; Minkoff, 2009; Nyden, Myren, & Gillberg, 2008; W. Pelham, Foster, & Robb, 2007). Adolescents with ADHD are at increased risk for a wide variety of functional impairments including lower academic and occupational achievement, lower rates of high school graduation, a higher incidence of teenage pregnancy and sexually transmitted infections, driving impairments, a greater number of legal charges/arrests, and higher rates of substance abuse (Babinski et al., 2011; Barkley, Fischer, Smallish, & Fletcher, 2004, 2006; Biederman et al., 2009; Bussing, Mason, Bell, Porter, & Garvan, 2010; Fletcher & Wolfe, 2009; Langley et al., 2010; Molina et al., 2007, 2008, 2009; Monuteaux, Mick, Faraone, & Biederman, 2010; Realmuto et al., 2009). Thus, ADHD in adolescence is not a benign condition and requires intervention.

Treatment.

The American Academy of Pediatrics and the American Academy of Child and Adolescent Psychiatry have treatment guidelines that specify how optimally to manage child and adolescent ADHD (American Academy of Pediatrics, 2001; Pliszka, 2007). Pharmacological management (i.e., stimulant medication) is the most widely explored and utilized form of treatment for children diagnosed with ADHD (DuPaul, Barkley, & Connor, 1998; Group, 1999; Solanto, Arnsten, & Castellanos, 2001). The use of stimulants appears to improve hyperactive symptoms, impulsivity, inattentiveness, and improve reaction time, task persistence, work productivity, and vigilance (DuPaul et al., 1998; Greenhill, Halperin, & Abikoff, 1999; Rapport & Kelly, 1993; Solanto et al., 2001).

Although the majority of children and adolescents respond favorably to stimulants, 20% to 30% of children and adolescents show no response (Swanson, McBurnett, Christian, & Wigal, 1995) or develop considerable enough side effects (e.g., weight loss, sleep disturbance, etc.) that the medication must be discontinued (Barkley, McMurray, Edelbrock, & Robbins, 1990; DuPaul et al., 1998; Findling, Short, & Manos, 2001; Gillberg et al., 1997).

Moreover, adherence to stimulant medications in ADHD wanes as a function of age (Charach, Ickowicz, & Schachar, 2004). The reason for the decline may be any of a number of factors including the following: (a) unlike in pediatric ADHD, where the parents are typically responsible for the child’s adherence to treatment, ADHD treatment in a high school students requires far more involvement of the patient himself or herself (Pliszka, 2007) and (b) stimulant misuse or diversion is another potential cause of less robust treatment adherence in adolescents with ADHD. Although more likely to happen in high school students with ADHD and comorbid conduct disorder or substance abuse (Gordon, Tulak, & Troncale, 2004; Thiruchelvam, Charach, & Schachar, 2001), stimulant misuse/diversion is a real possibility in any adolescent receiving stimulant therapy for ADHD (Faraone & Wilens, 2007; Wilens et al., 2008).

Aside from medication, other empirically validated psychosocial treatments for children with ADHD include behavioral parent training and behavioral school interventions (Lonigan, Elbert, & Johnson, 1998; W. E. Pelham, Wheeler, & Chronis, 1998), academic interventions such as task and instruction modifications, strategy training, peer tutoring, and computer-assisted instruction (DuPaul & Eckert, 1998) and combined behavioral-pharmacological interventions (Group, 1999; Klein et al., 1997). Most psychosocial interventions for children with ADHD focus helping the adult caretakers of the child with ADHD to better manage the child (Pliszka, 2007).

Compared with the vast amounts of research on psychosocial interventions in children, far less research has been conducted on the efficacy of psychosocial interventions for adolescents with ADHD. This is puzzling given that parents of adolescents prefer psychosocial intervention to medication, yet are more likely to have access to medication (Bukstein, 2004; Jensen et al., 1999; McLeod, Fettes, Jensen, Pescosolido, & Martin, 2007). Furthermore, parent satisfaction ratings are higher when parents received psychosocial treatment than when they received medication alone (MTA Collaborative Group, 1999b), suggesting that despite the lack of research attention, psychosocial treatments may have a place in the treatment of ADHD.

One of the most common psychosocial interventions in other forms of psychopathology is CBT. Used with both children and adults, CBT has a substantial research base supporting its efficacy for a wide variety of both medical and psychiatric conditions (especially internalizing disorders). For example, pediatric meta-analyses have documented medium-to-large effect sizes (d > 0.5) for CBT in the management of obsessive compulsive disorder (OCD; Watson & Rees, 2008) and depression (Chu & Harrison, 2007). Although CBT clearly has demonstrated efficacy in child and adolescent populations, it is generally accepted that CBT is more effective in adolescents relative to preadolescents (Holmbeck, Greenley, & Franks, 2003), implying that cognitive development may moderate the efficacy of CBT treatment.

CBT has not been found to be particularly beneficial for children with ADHD (Abikoff & Gittelman, 1985; Baer & Nietzel, 1991; Bloomquist et al., 1991; DuPaul, & Eckert, 1997; Dush, Hirt, & Schroeder, 1989). Although there is a database suggesting that CBT may be more efficacious for adults with ADHD (Safren et al., 2005, 2010), the literature on CBT for adolescent ADHD is scarce. In our exhaustive review of the literature, we were unable to find any data which have been published on the use of CBT in adolescent ADHD. Nonetheless, several adult ADHD treatment studies have employed CBT models as an intervention for reducing ADHD symptoms and associated impairments.

Ramsay and Rostain (2008) have also created a CBT program for adults with ADHD. This consists of 16 therapy sessions over a period of 6 months. The program focuses on developing and implementing coping strategies (e.g., organizational skills and time management) to manage ADHDrelated symptoms, as well as identifying and modifying dysfunctional thoughts and beliefs. In an open study by Ramsay and Rostain, adults with ADHD were treated with a combination of pharmacotherapy and CBT (Rostain & Ramsay, 2006). The findings suggested that the combined treatment approach was efficacious across both symptom and functional parameters. Overall, 67% of the patients were rated by clinicians as much or very much improved in ADHD symptom severity.

Philipsen and colleagues (2007) developed a group CBT approach for managing adult ADHD. Over 13 weekly 2-hr sessions, adults with ADHD participated in groups covering various topics related to ADHD. Philpsen et al. studied 72 adult participants with ADHD (29 females, 43 males; M age = 36 years), most of whom (81.9%) had comorbid psychopathology (Philipsen et al., 2007). Forty-eight adults with ADHD entered the study already receiving medication and 24 were not receiving pharmacotherapy. The results indicated that ADHD and depressive symptoms decreased, both with large effect sizes. Individuals with and without medication improved similarly. Only six adults with ADHD failed to complete the 13-week program (Philipsen et al., 2007). Others (Bramham et al., 2009) have similarly studied group CBT approaches to managing adult ADHD with comparable results.

Given this dearth in the literature as well as the clear need for the use of psychosocial interventions in adolescents with ADHD, we sought to assess the efficacy of a CBT treatment approach used in adult ADHD (Safren et al., 2005, 2010) in a sample of adolescents with ADHD.

Adolescence is a period marked by increased desire for independence and individuation from the family. Thus, rather than working primarily with parents (as exists in many child ADHD treatment paradigms), it makes clinical sense to involve adolescents in treatment. We believe that this provides rationale for using CBT and incorporating a direct focus on including the adolescent in treatment.

Our a priori hypothesis was straightforward: Adolescents with ADHD who received CBT would have a significant reduction in parent- and teacher-rated ADHD symptoms and a significant improvement in functional outcomes related to academic, social, and family domains. In a more exploratory fashion, we sought to examine the effect of ADHD subtype, gender, and psychiatric comorbidity on clinical outcomes in our treated sample.

Participants

Participants were recruited from the Adult ADHD Treatment and Research Program at State University of New York (SUNY)—Upstate Medical University. The “adult” program serves individuals high school age (9th-12th grade) and older. Referrals to the Adult ADHD program are primarily from primary care physicians (34%), educational settings (33%), and parents (20%). For the current study, only those adolescents currently in high school were included. High school students currently comprise approximately 45% of the total adult ADHD program population. The institutional review board approved the study and informed consent/assent was obtained from all participants.

Participating adolescents were selected from consecutive referrals to the Adult ADHD treatment program over a 4-year period (September 1, 2006-August 31, 2010). Although 127 adolescents were referred to the ADHD treatment program, only 82 met formal Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; APA, 1994) diagnostic criteria for ADHD (combined type: n = 40; inattentive type: n = 42). Adolescents with ADHD were primarily Caucasian (81%) with African American (13%) and Latino (3%) comprising the remaining participants. At the time of the evaluation, the mean age of the adolescent participants was 16.4 years (SD = 1.3 years; range = 14 years-18 years, 2 months) and consisted of more males (n = 51) than females (n = 31). There were no gender, χ² = 1.21, p = .543, or age, F(1, 80) = 0.56, p = .689, differences between participants with ADHD–combined and ADHD–inattentive types (see Table 1 for complete background data).

Of the 82 adolescents with ADHD, 10 adolescents/parents of adolescents declined to participate in the treatment. The three primary reasons for declining treatment were low adolescent interest in the treatment program (n = 7), lack of time (n = 6), and distance from the clinic (n = 5). The 10 adolescents with ADHD who declined treatment did not differ from the 72 adolescents with ADHD who elected to participate in the treatment on any demographic, clinical, or cognitive variable (p > .20).

Of the remaining 72 adolescents with ADHD who elected to participate in the treatment, 4 adolescents with ADHD failed to complete the entire CBT treatment protocol. Two of the 4 adolescents who failed to complete the protocol moved away from the area; the third adolescent was placed in a residential treatment facility, and the other adolescent refused to continue in therapy. These adolescents who failed to complete the treatment did not differ in age, gender, or other clinical variables (p > .20) from those 68 adolescents who did complete the entire protocol. All remaining data are from those 68 adolescents with ADHD who completed the treatment protocol.

As is common in ADHD, most adolescents (n = 37) met diagnostic criteria for comorbid oppositional defiant disorder. In addition, generalized anxiety disorder (n = 27), major depressive disorder (n = 24), substance abuse (n = 20), conduct disorder (n = 12), OCD (n = 12), and learning disabilities (n = 5) were the most common comorbid conditions. Most adolescents with ADHD had a comorbid condition; only 20% of adolescents had the single diagnosis of ADHD. Clinical-rated mean global assessment of functioning (GAF; APA, 2000) score was 52.2 (SD = 11.5), indicative of moderate to serious symptoms and impairment ().

All adolescents with ADHD who participated in the study also participated in concurrent pharmacotherapy. Fifty-nine of the adolescents were already prescribed stimulant (n = 54) or nonstimulant (atomoxetine; n = 5) medication at the time of referral. The remaining nine adolescents with ADHD began stimulant therapy shortly after the evaluation yet before CBT was initiated. Using a 3-item Likert-type scale (0 = poor, 1 = satisfactory, 2 = good), treatment adherence in the month prior to initiating the CBT program was estimated by parents to be between satisfactory and good (M = 1.35, SD = 0.45).

To increase the ecological validity of this study, no a priori decision was made to only include medication responders or medication nonresponders prior to initiating the CBT protocol. Regardless of medication responsiveness, all high school students who were referred to the Adult ADHD Treatment and Research Program and who met DSM-IV criteria for ADHD were invited to participate in the CBT trial.

Educationally, 61% of the adolescents with ADHD (n = 41) were served on some form of special education plan, generally an Americans with Disabilities Act (ADA) 504 plan (n = 27) or an individualized Education Plan (IEP; n = 14). Although there were no specific study prohibitions against making changes to the educational plan, no changes were made to the special education plans or services during the treatment phase.

All 68 adolescents with ADHD included in the current study had prior pharmacological treatment for ADHD. Just over half of the adolescents (n = 36) had previously also received some form of psychosocial intervention including parent training (n = 19), family therapy (n = 11) or individual therapy for the child (n = 18). At the time of participation in the CBT, however, no adolescents with ADHD were also receiving concurrent psychosocial interventions.

Measures

Psychiatric/behavioral.

Psychiatric assessments of adolescents relied on the Schedule for Affective Disorders and Schizophrenia, Epidemiologic version (K-SADS-E; Orvaschel & Puig-Antich, 1987). ADHD diagnoses were based on independent interviews with the parents/guardians (almost always a mother or stepmother) and direct interviews of adolescents. When the information from parent and adolescent was discrepant, the most severe report was utilized. For example, when one reporter endorsed a symptom and the other did not, we counted the symptom as present. For every ADHD diagnosis, information was also gathered regarding the ages at onset and offset of symptoms, functional impairments, and treatment history. Diagnoses were considered positive if, based on the interview results, DSM-IV criteria were unequivocally met to a clinically meaningful degree. Diagnoses were made for two points in time: lifetime and current (past month). All adolescents had positive histories for both child ADHD (during elementary school) and current ADHD (during high school). No adolescents received de novo ADHD diagnoses; all had preexisting diagnoses which were confirmed using the K-SADS-E. All K-SADS-E interviews were conducted by the first author, a licensed clinical psychologist who has extensive experience with the K-SADS-E. No interrater reliability statistics were computed. In addition to ADHD diagnoses, the K-SADS-E was employed to assess for psychiatric comorbidity, both past and current.

The Behavior Assessment System for Children–2nd edition (BASC-2; Reynolds & Kamphaus, 2006) was administered to assess adolescent behavioral functioning. The 126-item BASC parent rating scale was administered to parents, whereas the 138-item BASC teacher rating scale was administered to the teacher who “knew the adolescent the best.” Adolescents completed the Self-Report of Personality (SRP) form. The BASC scales yield composite T-scores on four dimensions: Externalizing Problems, Internalizing Problems, Behavior Symptoms, and Adaptive Skills. The teacher rating scale includes an additional school Problem Composite. Scores >65 indicate clinical significance. The BASC possesses adequate psychometric properties; validity has been demonstrated by associations with other measures of adolescent behavior and intervention studies (Reynolds & Kamphaus, 2006).

The ADHD–Rating Scales (ADHD-RS; DuPaul, Power, & Anastopoulos, 1998) was used as a measure of parentand teacher-rated ADHD symptoms. The ADHD-RS is an 18-item questionnaire requiring the respondent to rate the frequency of the DSM-IV symptoms of ADHD using a 4-point Likert-type scale (0 = never or rarely, 1 = sometimes, 2 = often, 3 = very often). This measure yields a total score and two subscale scores (Inattention and Hyperactivity-Impulsivity). The reliability and validity of the ADHD-RS are adequate (DuPaul et al., 1998). All adolescents who met Diagnostic and Statistical Manual of Mental Disorders (4th ed., text rev.; DSM-IV-TR; ) diagnostic criteria for ADHD had an ADHD-RS score >1.5 SD above age and gender norms. Raw scores for all three scales were used as dependent measures.

Finally, the clinician-rated GAF () was also used as a measure of clinical severity and functioning.

Cognitive.

The Wechsler Abbreviated Scale of Intelligence (WASI; Wechsler, 1999), which comprises four subtests: Vocabulary, Block Design, Similarities, and Matrix Reasoning, was administered to obtain a Full-Scale IQ estimate.

Functioning.

In addition to the adaptive functioning scales included in the BASC-2, functional outcomes were assessed via parent and teacher report on the Impairment Rating Scale (IRS; Fabiano et al., 2006). The Parent IRS consists of 7-items (relationship with peers, relationship with siblings, relationship with parents, academic progress, self-esteem, influence on family functioning, and overall impairment). The Teacher IRS teacher version has six domains (relationship with peers, relationship with teacher, academic progress, self-esteem, influence on classroom functioning, and overall impairment). Parents and teachers are instructed to rate the severity of the adolescents problems in each domain and the need for treatment using a 7-point Likert-type scale (0 = no problem/definitely does not need treatment, 6 = extreme problem/definitely needs treatment). The IRS has been demonstrated to have good validity and reliability (Fabiano et al., 2006).

In addition, several ecologically valid measures of realworld functioning were obtained. These measures included cumulative grade point average and number of school absences/tardies.

Procedures

During the evaluation, all of the dependent measures utilized in the study were administered and scored. All 82 adolescents diagnosed with ADHD had a feedback session in which the diagnosis or diagnoses were described. During this meeting, all 82 adolescents and their families were encouraged to consider a combined treatment approach of CBT and pharmacotherapy. For the 72 adolescents who elected to participate in CBT, a therapy session was scheduled and treatment was initiated.

Teachers were mailed rating scales during the diagnostic process as well as at the end of treatment (on average, 5 months later). Teachers were blind to the child’s CBT participation.

CBT intervention

The intervention used in the study was a downward extension of the Safren et al. (2005) program for adults with ADHD who have symptoms unresolved by medication. The Safren et al. CBT model involves components of motivational interviewing, practice, review, and repetition of learned skills and is comprised of three core modules in addition to three optional modules. In the current study, all participants received all six modules (core plus optional modules) and all sessions were led by a doctoral-level clinical psychologist.

The first four-session core module involved psychoeducation about ADHD as well as training in organization and planning skills. The second 3-session core module focused on the adolescent learning skills to reduce distractibility. The third core module used cognitive restructuring strategies described by Beck (1995) modified to account for an ADHD population. This module varied between two and five sessions (M = 3.1 session, SD = 1.1 sessions) depending on the particular adolescent. The Safren et al. (Safren et al., 2005) optional modules were also completed by the adolescents with ADHD. These modules included four sessions focused on reducing procrastination, improving communication skills (e.g., reducing interruptions, improving active listening, etc.) and improving anger/frustration management.

Handouts describing the topics covered were provided to the parent/adolescent at the conclusion of each session. The handouts were provided in an attempt to encourage generalizability of the strategies to the real world.

Parent involvement.

Parents and guardians were actively included in the assessment and CBT treatment. Parents were in the therapy room during the first seven sessions (Core Modules 1 and 2) as well as the one optional procrastination module. Parents were not in therapy room during Core Module 3 or the communication skills or anger management optional modules. When parents were present in the therapy room, parents were instructed explicitly that they were present only to “improve generalizability of the strategies to the real world.” The therapist directed the vast majority of the session toward the adolescent with time for the parent to ask questions/seek clarification at the end of the 50-min session. When parents were not included in the therapy room, parents were informed of the topics discussed, yet not the specific content of the topics.

Adolescent treatment adherence

CBT adherence.

All adolescents completed all modules (range = 13-16 sessions). Attendance was generally high with 55% of the adolescents never missing an appointment. The remaining 45% of adolescents missed between 1 and 3 appointments (missed appointments M = 1.4 appointments, SD = 0.9). Given that most of the adolescents were not yet of driving age or did not yet have their driver’s permit, parental attendance at the sessions was highly consistent with adolescent attendance.

Medication adherence.

In the first session, parents were instructed to monitor their child’s medication adherence. At the beginning of each session, parents were asked to rate their child’s adherence to the medication regimen during the last week (or 2 weeks if sessions were spaced 2 weeks apart) on a 3-item Likert-type scale (0 = poor, 1 = satisfactory, 2 = good). Mean treatment adherence during the CBT program was estimated by parents to be good (M = 1.68, SD = 0.29).

Planned analyses

The study used a standard pre- and posttreatment design. Paired samples t tests were utilized to assess change. Scores at the pre- and posttreatment periods were calculated for outcome variables. Outcome variables consisted of both symptom measures (BASC-2 and ADHD-RS) as well as functioning measures (BASC-2, IRS, several ecologically real-world measures). Cohen’s effect size (Cohen’s d) is also reported for pre- and posttreatment analyses.

Following these analyses, and in a more exploratory fashion, the role of ADHD subtypes (ADHD–inattentive vs. ADHD–combined), gender, and psychiatric comorbidity was investigated using repeated measures ANCOVA using pretreatment scores as the covariate. In the case of psychiatric comorbidity, many adolescents had 3+ comorbid conditions. In these cases, and only for the purpose statistical analyses, an a priori decision was made to consider the most impairing comorbid condition as the psychiatric comorbidity. This decision was made on the basis of clinical judgment and Parent BASC-2 scores. For example, if the Parent BASC-2 had higher T-scores for anxiety and the clinician agreed that the adolescent was more anxious than oppositional, depressed, and so on, then the anxiety disorder diagnosis was considered as the most impairing comorbid condition.

Due to the high number of comparisons, we applied the standard Bonferroni correction method to all significant analyses. Only those results significant at the p ≤ .008 level can be considered statistically significant.

Treatment efficacy

As demonstrated in Table 1, pretreatment and posttreatment effects were found for several symptom and functional variables. The largest effect sizes were observed for number of weekly missed classes, school tardies, stimulant medication doses, parent-reported externalizing behaviors, parent-reported inattention symptoms, and teacher-reported inattention symptoms. Across all variables, participants had lower numbers at posttreatment relative to pretreatment.

As also noted in Table 1, the adolescents reported less pre- and posttreatment changes than parents and teachers (see Table 1 for complete results).

Exploratory analyses

In a more exploratory fashion, ADHD subtype, gender, and psychiatric comorbidity were considered as possible moderators of treatment outcomes. Using an ANCOVA with pretreatment scores as the covariate, no statistically significant interactions emerged between ADHD subtype or gender, time, and any of the dependent variables (p > .20). Thus, our adolescent CBT protocol does not appear to be more effective with either of the subtypes or as a function of gender.

In regards to psychiatric comorbidity, four groups were created: ADHD only (n = 14), ADHD + ODD/CD (n = 25), ADHD + ANX (n = 15), and ADHD + DEP (n = 14). As demonstrated in Table 2, Time × Psychiatric Comorbidity (F degrees of freedom = 3, 64) interactions were found on several parent-, teacher-, and adolescent-report measures of behavioral symptoms and functioning. Once Bonferroni corrections were applied, however, only several variables maintained statistical significance. Scheffé post hoc tests revealed that across all outcome variables, the group of adolescents with ADHD + ODD was rated by parents and teachers as benefiting less from the CBT intervention. Scheffé post hoc tests also revealed that the adolescents with ADHD + ANX and the adolescents with ADHD + DEP improved more than the ADHD only group on several variables.

ADHD is a well-studied psychiatric disorder in children. However, far less is known about adolescents with ADHD. Given that the majority of children with ADHD become adolescents with ADHD, this is an unfortunate dearth in the extant literature, especially when one considers the economic toll created by untreated or poorly treated ADHD (Biederman & Faraone, 2006). The most common of these treatments, pharmacotherapy, remains a mainstay of pediatric ADHD treatment. However, some children with ADHD do not respond to stimulants or cannot tolerate the stimulants. In addition, in adolescence, stimulant misuse and diversion can become challenges. All of the preceding suggest that there remains a clear need for effective psychosocial interventions in adolescents with ADHD.

These data suggest that a downward extension of an empirically validated adult ADHD CBT protocol (Safren et al., 2005) can benefit some adolescents with ADHD. Treatment outcomes were similar for males and females with ADHD as well as those with ADHD–inattentive type and ADHD–combined type. Nonetheless, those adolescents with comorbid ODD or CD benefited less than the adolescents with only ADHD or with comorbid anxiety disorders or depression. Adolescents with ADHD only, ADHD + ANX, and ADHD + DEP were rated by parents and teachers as improved across several symptoms and, more importantly, functional parameters including several ecologically valid “real-world” measures such as grades and attendance.

In addition, over the course of the CBT, lower doses of medication were needed to maintain the adolescent’s functional improvements. To our knowledge, this is the first psychosocial study of adolescents or adults with ADHD that has tracked medication doses as a function of treatment time. Given the concerns that some primary care physicians have with prescribing stimulant medications to adolescents (Wilens et al., 2008), the use of a CBT protocol may help to reduce the doses of the medications needed to improve functioning.

Despite the gains made by the adolescents with ADHD only, ADHD + ANX, and ADHD + DEP, many of these adolescents did not normalize their functioning and remained both symptomatic and functionally impaired (as rated by teachers and parents) in at least one domain. It is possible that the lower doses of medications were partly responsible for this finding. Alternatively, it is also possible that these adolescents with ADHD are continuing to function somewhat below average as a function of the reputation that each has developed over time with their parents, teachers, and peers. Future research should consider how best to explain why despite improvements, most adolescents with ADHD continued to function suboptimally.

Other research to consider in the future includes the role of the parents in the therapy session. As described earlier, parents were included for much of the direct work with the adolescent. Although this may have helped to improve generalizability of the strategies, it may also have worked against those adolescents with ODD or CD who may have come to view the therapist as simply another adult in their life who is “trying to fix” him or her.

Likewise, future research should also consider how best to explain that the parents and teachers reported far more appreciable improvements than the adolescents. A possible explanation for this finding include adolescent optimistic bias, often referred to as a positive illusory bias in the ADHD literature (Mikami, Calhoun, & Abikoff, 2010); across all assessed parameters, the adolescent with ADHD rated himself or herself as less symptomatic and higher functioning than the teacher and parent. Thus, by having an inflated selfperception, the adolescents had less room for improvement, hence lower effect sizes. Future research should consider the impact that adolescent self-perception has on treatment outcomes as well as the validity of such self-perceptions.

The current research project represents an initial foray toward furthering study of psychosocial interventions in adolescents with ADHD. These data represent data amassed from an ongoing clinical program designed to treat adolescents with ADHD. These data are not from an artificial research study which may not approximate what clinicians can manage in an Health Maintenance Organization (HMO)-driven reality. As such, these clinical data need to be considered in light of several methodological limitations that may not have been present in a tightly controlled research study. First and foremost, our treatment group was self-selected for accepting CBT therapy. This group may differ systematically in traits (e.g., motivation) that could have confounded our results. Second, our outcome assessments were not blinded to group status. Finally, given that the study was conducted in a natural clinical setting, with limited resources, rather than a designated research clinic, interrater reliability on psychiatric diagnoses and treatment integrity were not formally measured by the investigators. Given these considerations, it is essential that a well-controlled efficacy trial be conducted to compare CBT therapy with an appropriate placebo control.

Despite our methodological limitations, these data represent the clinical reality in which most mental health professionals working with adolescents with ADHD exist. For example, few parents of functionally impaired adolescents with an ADHD are willing to have their child wait for treatment for several months simply to participate in a research study. In addition, although consultation with schools is clearly an important component of pediatric mental health treatment, having the resources to be able to provide ongoing in-service training to teachers as was done in the MTA study (MTA ) is also likely beyond the means of the clinician working in a managed care environment.

Although this work must be considered preliminary due to its methodological limitations, it suggests that large, well-controlled trials of CBT for adolescent ADHD are warranted. Although CBT has not been extensively used within the adolescent ADHD domain, there exists a vast literature in the adolescent depression and anxiety domains which detail the effectiveness of CBT (Chu & Harrison, 2007; Watson & Rees, 2008). This is promising yet awaits replication in more carefully controlled trials before similar conclusions can be drawn about CBT in adolescent ADHD.

In addition, the long-term maintenance of these treatment gains remains an open question. Future research should continue to follow adolescents with ADHD over time to gauge the impact that the interventions have as a function of time. Although short-term gains are promising and encourage future research, the long-term maintenance of this CBT intervention is presently unclear.